Uncertainty in Calculating Hydrological Drought Index Using Gamma Distribution (Case Study: Urmia Lake Basin)

This study analyzes the Uncertainty arising from the assumption of fitting the Gamma distribution to the time series of discharge data in calculating the Streamflow Drought Index (SDI). Using the time series of monthly and annual discharge data in 30 hydrometric stations located in the Urmia Lake Basin, the efficiency of Gamma distribution in fitting discharge data was investigated. The results showed that the Gamma distribution in the fitting of the discharge data series was recognized as the superior distribution in only 1.5% of the total cases. According to Kolmogorov-Smirnov test, a significant difference was observed between the experimental distribution of discharge data and the distribution of Gamma in 25% of cases, which increases to about 57% in the first and last months of the wet year. In contrast, the Wakeby distribution was recognized as the best distribution in fitting the discharge data series in 31% of cases. The highest displacement of hydrological drought classes was observed with the use of Wakeby distribution compared to Gamma distribution in the monthly scale of Ghabeghloo station in August (96%) and in the annual scale of Ghasemloo station (54%). It was observed that by calculating SDI based on Wakeby distribution compared to Gamma distribution, the highest frequency of displacement in annual and monthly scales is related to normal class and is equal to 25% and 43%, respectively.